From owner-freebsd-hackers Mon Oct 6 07:44:07 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id HAA02397 for hackers-outgoing; Mon, 6 Oct 1997 07:44:07 -0700 (PDT) (envelope-from owner-freebsd-hackers) Received: from unix.tfs.net (root@unix.tfs.net [199.79.146.60]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id HAA02389 for ; Mon, 6 Oct 1997 07:44:03 -0700 (PDT) (envelope-from jbryant@argus.tfs.net) Received: from argus.tfs.net (node21.tfs.net [207.2.220.21]) by unix.tfs.net (8.8.5/8.8.5) with ESMTP id JAA19150; Mon, 6 Oct 1997 09:42:35 -0500 Received: (from jbryant@localhost) by argus.tfs.net (8.8.7/8.8.5) id JAA03872; Mon, 6 Oct 1997 09:43:55 -0500 (CDT) From: Jim Bryant Message-Id: <199710061443.JAA03872@argus.tfs.net> Subject: Re: Building parallel "Beowulf-style" supercomputers with FreeBSD In-Reply-To: from "Ron G. Minnich" at "Oct 6, 97 09:09:36 am" To: rminnich@Sarnoff.COM (Ron G. Minnich) Date: Mon, 6 Oct 1997 09:43:55 -0500 (CDT) Cc: freebsd-hackers@freebsd.org Reply-to: jbryant@tfs.net X-Windows: R00LZ!@# MS-Winbl0wz DR00LZ!@# X-Operating-System: FreeBSD 2.2.2-RELEASE #0: Wed Jul 9 01:01:24 CDT 1997 X-Mailer: ELM [version 2.4ME+ PL31H (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-hackers@freebsd.org X-Loop: FreeBSD.org Precedence: bulk In reply: > On Sat, 4 Oct 1997, Douglas Carmichael wrote: > > Is there any way one can build large "Beowulf-like" clusters > >with FreeBSD and its SMP functionality? > > Anyone built such a beast? > > [. . .] > > not to the level that differentiates it in a big way. Beowulf is a special > instance of a general idea, the compute cluster. People have been > building compute clusters for almost ten years, some of them using up to > 300 processors. (some names: Fermilabs; SCRI; an interesting site in > Alabama of all places; Sun Labs [Mica]. All built 128+ processor > clusters at least five years ago). A cluster based on FreeBSD would > I suppose have to have a different name, but we've done a lot of computing > over the last three years with such a cluster: works fine. Was not the first at Sandia, using VAXen? '88? Hmmmmm.... Parallel Virtual Machines or some name like that? I recall that on the test apps, it outperformed a X/MP or Y/MP... Do the programs have to be hand-designed, or is there a vectorizing compiler available? I have a copy of the PVM[?] stuff somewhere, and remember that it required hand coding the whole thing... Or is this just SMP stuff... Hmmm... A roomful of PPros or P-II's, vectorizing compiler.... [oops, I'm drooling]... Of course, I always thought that this was simply an extension of the transputer concept, except using standard networks, and standard computers instead. Nothing wrong with this. jim -- All opinions expressed are mine, if you | "I will not be pushed, stamped, think otherwise, then go jump into turbid | briefed, debriefed, indexed, or radioactive waters and yell WAHOO !!! | numbered!" - #1, "The Prisoner" ------------------------------------------------------------------------------ Inet: jbryant@tfs.net AX.25: kc5vdj@wv0t.#neks.ks.usa.noam grid: EM28pw voice: KC5VDJ - 6 & 2 Meters AM/FM/SSB, 70cm FM. http://www.tfs.net/~jbryant ------------------------------------------------------------------------------ HF/6M/2M: IC-706-MkII, 2M: HTX-212, 2M: HTX-202, 70cm: HTX-404, Packet: KPC-3+